Lecture 5: Measures of Association

what does it mean if relative risk is technically greater than or less than 1, but is very close to 1.0

the actual effect of risk is negligible

what does it mean if attributable risk = 0

the level of risk between exposed and unexposed is the same. an exposure makes no difference to the risk of developing disease

what is absolute risk

the risk of developing a disease over time

if attributable risk between smoking and lung cancer is 10.6 per 1000 people, what does that mean in words

there are 10.6 excess cases of lung cancer per 1000 persons that are attributable to smoking

if the odds ratio between asbestos exposure and mesothelioma was 2.85 what does that mean

those that have mesothelioma have 2.85 greater odds that they were exposed to asbestos compared to those without mesothelioma

6000 participants in a study, 3000 with bladder cancer and 3000 without. of those with bladder cancer, 1293 used artificial sweeteners at some point, 1707 never had. of those without bladder cancer, 855 had used sweeteners, 2145 had not. odds ratio = 1.9 interpret odds ratio

those who have bladder cancer as 1.9 times as likely to have used sweetener than those that do not have cancer

1000 participants in a study where 400 were smokers and 600 were not, 50 smokers developed bladder cancer, 15 nonsmokers developed cancer. relative risk = 5 interpret relative risk

those who smoke are 5x as likely to develop bladder cancer as those who do not smoke

what is a confidence interval

used to express statical precision around the point estimate

a 200 participants study used 100 participants in the treatment group and 100 in the control group. 10% of the control group experienced neuropathy and 3% of treatment group experienced neuropathy. what is absolute risk reduction

(10/100) - (3/100)

4000 patients were recruited where 2000 had mesothelioma. 1100 of the mesothelioma patients were also exposed to asbestos and 600 without the disease were also exposed. what would the odds ratio be

(1100x1400) / (900x600) = 2.85

6000 participants in a study, 3000 with bladder cancer and 3000 without. of those with bladder cancer, 1293 used artificial sweeteners at some point, 1707 never had. of those without bladder cancer, 855 had used sweeteners, 2145 had not. what is the odds ratio

(1293x2145) / (855x1707) = 1.9

1000 participants in a study where 400 were smokers and 600 were not, 50 smokers developed bladder cancer, 15 nonsmokers developed cancer. what is attributable risk

(50/40) - (15/600) = .1

1000 participants in a study where 400 were smokers and 600 were not, 50 smokers developed bladder cancer, 15 nonsmokers developed cancer. what is relative risk

(50/400) / (15/600) = 5

8000 participants were included in a study, 3000 were smokers, and 84 of those developed CHD. of the non smokers, 87 developed the disease. what is the relative risk

(84/84+2916) / (87/87+4913) = 1.61

what is the 2x2 calculation for attributable risk

(a/a+b) - (c/c+d)

what is the 2x2 table calculation for relative risk

(a/a+b) / (c/c+d)

what is the 2x2 calculation of odds ratio

(a/c) / (b/d) or ad/bc

1000 participants in a study where 400 were smokers and 600 were not, 50 smokers developed bladder cancer, 15 nonsmokers developed cancer. attributable risk is .1 interpret attributable risk

.1 cases of bladder cancer can be attributed to smoking

a 200 participants study used 100 participants in the treatment group and 100 in the control group. 10% of the control group experienced neuropathy and 3% of treatment group experienced neuropathy. what is relative risk reduction

1 - [(3/100) / (10/100)]

what is the usual confidence interval

95%

what is the measure of association used by experimental study designs

absolute risk reduction relative risk reduction number needed to treat number needed to harm

what is the mathematical difference between attributable risk and relative risk

attributable is subtraction relative is division

a cohort study about flu vaccine and influenza was performed on 100 people. 65 people were exposed to the vaccine and 11 of those developed flu. 35 people were not exposed to the vaccine and 25 of those developed flu. what would the 2x2 table look like

box A - 11 box B - 54 total exposed - 65 box C - 25 box D - 10 total not exposed - 35 total flu - 36 total with no flu - 64 participants total - 100

4000 patients were recruited where 2000 had mesothelioma. 1100 of the mesothelioma patients were also exposed to asbestos and 600 without the disease were also exposed. what would the 2x2 box look like

box A - 1100 box B - 600 exposed total - 1700 not exposed total - 2300 box C - 900 box D - 1400 disease total - 2000 not disease total - 2000 participants total - 4000

6000 participants in a study, 3000 with bladder cancer and 3000 without. of those with bladder cancer, 1293 used artificial sweeteners at some point, 1707 never had. of those without bladder cancer, 855 had used sweeteners, 2145 had not. what is 2x2 table

box A - 1293 box B - 855 exposed total - 2148 box C - 1707 box D - 2145 not exposed total - 3852 disease total - 3000 not diseased total - 3000 total participants 60000

a 200 participants study used 100 participants in the treatment group and 100 in the control group. 10% of the control group experienced neuropathy and 3% of treatment group experienced neuropathy. what would the 2x2 table look like

box A - 3 box B - 97 exposed total - 100 box C - 10 box D - 90 not exposed total - 100 disease total - 13 not diseased total - 187 participant total - 200

1000 participants in a study where 400 were smokers and 600 were not, 50 smokers developed bladder cancer, 15 nonsmokers developed cancer. what is 2x2 table

box A - 50 box B - 350 exposed total - 400 box C - 15 box D - 585 not exposed total - 600 disease total - 65 not diseased total - 935 total participants - 1000

8000 participants were included in a study, 3000 were smokers, and 84 of those developed CHD. of the non smokers, 87 developed the disease. what does the 2x2 table look like

box A - 84 box B - 2916 exposed total - 3000 box C - 87 box D - 4913 non exposed total - 5000 disease total - 171 nondisease total - 7829 total participants - 8000

what study design uses odds ratio

case control study

what study often uses relative risk

cohort study

what is relative risk

compares the incidence between exposed and not exposed

what is the size of the confidence interval if there is increased certainty about the true effect size

confidence interval becomes more narrow as certainty increases

what is the mathematical basis for absolute risk reduction

control event rate - experimental event rate or (c/c+d) - (a/a+b)

what is the mathematical basis of relative risk reduction

control event rate - experimental event rate or 1- relative risk

what is attributable risk

difference in risk between exposed and not exposed

what does it mean if odds ratio = 1

disease is no associated with exposure

what are the columns of a 2x2 table based on

disease status

what are 2x2 tables

display dichotomies of exposed vs not exposed and disease vs no disease

what is a point estimate

effect size that is observed in a study

what is absolute risk reduction

excess risk associated with exposure/intervention as compared with the control

what does it mean if relative risk = 1

exposure is not associated with disease

what are the rows of a 2x2 table based on

exposure status

what does it mean if attributable risk is >0

exposure to the risk is harmful there is increase in disease in exposed group

what does it mean if attributable risk is <0

exposure to the risk is protective there is a reduction in disease in the exposed group

what does it mean if relative risk is >1

harmful exposure is associated with an increased risk of disease

what does it mean is odds ratio is >1

harmful increased odds of having the disease and exposure

what does attributable risk show

how much of an outcome/disease could be eliminated with better exposure control highlights effect of public health measures

what is the most basic expression of risk

incidence

what term is absolute risk interchangeable for

incidence

what is the mathematical basis of attributable risk

incidence in exposed - incidence in unexposed

what is the odds ratio

incidence is not available so we use an indirect measure of risk

what does it mean if a confidence interval includes 1.0 in its range

it crosses the null value so the data is insignificant and can be ignored

what are measures of association

measurements between exposure and disease that are used to compare risk

what does odds ratio compare

odds of a case being exposed to the odds that a control was exposed

what is the measure of association used by case control studies

odds ratio

what are case participants

people with disease

what are control participants

people without disease

what is relative risk reduction

proportion of risk reduction attributable to the exposure/intervention as compared to the control

what does it mean if odds ratio is <1

protective decreased odds of having the disease and exposure

what does it mean if relative risk is <1

protective exposure is associated with a decreased risk of disease

what is the mathematical basis for relative risk

ratio of risk associated with exposure to risk associated to no exposure incidence in exposed divided by incidence in unexposed

what is the measure of association used by cohort studies

relative risk attributable risk

what is another term for attributable risk

risk difference

if relative risk of smoking vs non smoking and CHD vs no CHD is 1.61, what does that mean in words

smokers were 1.61 times as likely to develop CHD compared to non-smokers